G.C. Ballintijn

The distributed ASCI Supercomputer project

The Distributed ASCI Supercomputer (DAS) is a homogeneous wide-area distributed system consisting of four cluster computers at different locations. DAS has been used for research on communication software, parallel languages and programming systems, schedulers, parallel applications, and distributed applications. The paper gives a preview of the most interesting research results obtained so far in the DAS project.

Algorithmic Design of the Globe Wide-Area Location Service

We describe the algorithmic design of a worldwide location service for distributed objects. A distributed object can reside at multiple locations at the same time, and offers a set of addresses to allow client processes to contact it. Objects may be highly mobile like, for example, software agents or Web applets. The proposed location service supports regular updates of an objects set of contact addresses, as well as efficient look-up operations. Our design is based on a worldwide distributed search tree in which addresses are stored at different levels, depending on the migration pattern of the object. By exploiting an objects relative stability with respect to a region, combined with the use of pointer caches, look-up operations can be made highly efficient.

Characterizing Internet performance to support wide-area application development

To help in our wide-area application development, we have done an informal study of the relation between wide-area latency, the number of routers, and geographical distance between Internet sites. We did this by performing ping and traceroute measurements between 19 sites distributed across the globe. Contrary to our expectation there is almost no correlation between distance, latency, and number of routers in the current Internet. To help further studies of these characteristics the problems of this pilot project are also described.

Achieving scalability in hierarchical location services

Services for locating mobile objects are often organized as a distributed search tree. The advantage of this approach is that the service can scale as a system grows in size and number of objects. However, a potential problem is that high-level nodes may become a bottleneck affecting the scalability of the service. A traditional solution is to also distribute the location information managed by a single node across multiple machines. We introduce a method that radically applies distribution of location information such that the load is evenly balanced across all machines that form part of the implementation of the service, while at the same time exploiting locality.

Mechanisms for effective caching in the Globe location service

Globe is a wide-area distributed system that supports mobile objects. To track and locate objects, we use a worldwide distributed location service, implemented as a search tree.An object registers its current position by storing its address in a nearby leaf node of the tree. This knowledge propagates up to the top of the tree, so every object can be found from the root. Remote objects can cache the location of an object. However, if the object moves, the cache entry is no longer valid.In this paper, we show how caching can be made to work effectively even in the presence of mobile objects.